Physics 251 Solution Set 1 Spring 2017

Transcription

1 Physics 5 Solution Set Spring 07. Consider the set R consisting of pairs of real numbers. For (x,y) R, define scalar multiplication by: c(x,y) (cx,cy) for any real number c, and define vector addition and multiplication as follows: (a) Is R a group? (x,y )+(x,y ) (x +x,y +y ), () (x,y ) (x,y ) (x x,y y ). () It is straightforward to check the group axioms and show that R is a group under addition [as defined in eq. ()]. R is not a group under multiplication. For example, (0,0) does not possess a multiplicative inverse. (b) Is R a field? R is not a field. Recall that all elements of a field, excluding the additive inverse, must possess a multiplicative inverse. In the case of R, the additive inverse is (0,0). However, for any x 0 and y 0, (x,0) and (0,y) also do not possess multiplicative inverses. (c) Is R a linear vector space (over R)? It is straightforward to check the axioms that define a linear vector space and show that R is a linear vector space over R. (d) Is R a linear algebra (over R)? It is straightforward to check the axioms that define a linear algebra and show that R is a linear algebra, where the vector multiplication law is given by eq. ().. Consider the following two groups: T { proper rotations that map a regular tetrahedron into itself }, T d { proper rotations and reflections that map a regular tetrahedron into itself }. Show that the following isomorphisms are valid: T A 4 and T d S4. A regular tetrahedron consists of four equilateral triangles with four vertices (labeled by the integers,,3,4) as shown in Figure (a). The symmetry operations, consisting of proper rotations [shown in Figure (b) and (c)] and reflections through planes that pass through two of the four tetrahedron vertices, have the effect of permuting the four vertices.

2 Figure : Proper rotations that map a regular tetrahedron into itself. Consider first the proper rotations. One can rotate by 0 in a clockwise or counterclockwise fashion about an axis through vertex, denoted by τ in Figure (b). This has the effect of permuting the vertices,3,4. All in all, one can perform clockwise or counterclockwise rotations about any one of the four axes [denoted by τ i, i,,3,4 in Figure (b)]. In each case, three of the four vertices are permuted. One can describe each rotation by an element of the permutation group. Using cycle notation, the eight rotation operations described above correspond to the three-cycles of the permutation group S 4. In cycle notation, these are: (3),(3),(4),(4),(34),(43),(34),(43). (3) In addition, one can rotate by 80 about the axes that are denoted by σ i, i,,3 in Figure (c). For example, performing a 80 about σ interchanges vertices and and likewise interchanges vertices 3 and 4. Thus, the three possible rotation operations described above correspond to permutations that are the product of two disjoint transpositions. In cycle notation, these are: ()(34),(3)(4),(4)(3) (4) Finally, the identity element corresponds to performing no rotation (or reflection). This completes the enumeration of all possible proper rotations that map a regular tetrahedron into itself. We conclude that T {e,(3),(3),(4),(4),(34),(43),(34),(43),()(34),(3)(4),(4)(3)}. (5) These twelve elements correspond to the even permutations of the four vertices. Therefore, it follows that T A 4, where A 4, the alternating group of four objects, is the subgroup of the permutation group S 4 that consists of the even permutations of four objects. Even permutations can be expressed as the product of an even number of transpositions. Moreover, as shown in class, an n-cycle can be written as the product of n transpositions. Hence, it follows that 3-cycles must be even permutations. Thus, eq. (5) exhausts all the possible even permutations of four objects.

3 Figure : Consider the bisectors of two of the equilateral triangles that are faces of the tetrahedron. The two bisectors meet at a point. The other ends of the two bisectors are connected by one of the tetrahedron edges. This defines one of six possible reflection planes that pass through two of the four vertices of the tetrahedron. One can also consider six possible reflection planes, one of which is illustrated in Figure. Note that each reflection plane passes through two of the four possible tetrahedron vertices. There are 4!/(!!) 6 ways of choosing the two vertices. When a reflection through one of these planes is carried out, the tetrahedron is mapped into itself. The two vertices that are located on the reflection plane are unaffected by the reflection, whereas the other two vertices are interchanged. These correspond to the transpositions of S 4, (),(3),(4),(3),(4),(34). (6) One can also combine any one of these reflections with a proper rotation. It is sufficient to consider one reflection (e.g., the reflection that interchanges vertices and ). There are six new permutations that can be produced: ( 3 4) ( 3 4), ( 4 3) ( 4 3), ( 3 4) ( 3 4 ), ( 4 3) ( 4 3 ), ( 3)( 4) ( 4 3), ( 4)( 3) ( 3 4). 3 4

4 Combining the reflection that interchanges vertices and with the five remaining rotations listed in eqs. (3) and (4) yields the five remaining transpositions, ( 3) ( 3), ( 3 ) ( 3), ( 4) ( 4), ( 4 ) ( 4), (3 4) (3 4) Thus, all the odd permutations of the four tetrahedron vertices can be realized by either a single reflection or a reflection followed by a rotation. One can easily check that two successive symmetry operations of the tetrahedron are in one-to-one correspondence with the multiplication table of S 4. Thus, we conclude that T d S4. 3. Consider the possibility that a set G of n n matrices forms a group with respect to matrix multiplication. (a) Prove that if G is a group and if one of the elements of G is a non-singular matrix then all of the elements of G must be non-singular matrices. Conclude that all the elements of G are either non-singular matrices or singular matrices. Let G {A 0,A,A,...} be a group of n n matrices, where e A 0 is the group identity element. First, suppose that the identity element A 0 is a non-singular matrix, in which case deta 0 0. Then consider A i B i A 0, for i 0 (no sum over i), (7) where B i is the group inverse of A i. Taking the determinant of both sides of eq. (7), it follows that deta i 0 and detb i 0. That is, A i is a non-singular matrix for all i. Hence, if the identity element is a non-singular matrix, then all the elements of G are non-singular matrices. Next, suppose that the identity element A 0 is a singular matrix, in which case deta 0 0. Since A 0 is the group identity element, it follows that A i A 0 A i, for any i 0. (8) Taking the determinant of both sides of eq. (8), it follows that deta i 0 for all i. Hence, if the identity element is a singular matrix, then all the elements of G are singular matrices. The group G may be a discrete or continuous group of matrices. 4

5 REMARKS:. In the case where all elements of G are non-singular matrices, then we can multiply both sides of eq. (8) by the matrix inverse A i to conclude that A 0 n n, where n n is the n n identity matrix. In the case where all the elements of G are singular matrices, then A 0 cannot be the identity matrix (since n n is non-singular).. One can shorten the above proof by proving directly that if any element of G is singular then all elements of G are singular. Suppose x G is a singular matrix, in which case detx 0. Consider any other element y G where y x. Then by writing y x(x y), (9) and taking the determinant of both sides of eq. (9), it follows that dety detx det(x y) 0. Hence, if any element of G is a singular matrix then all elements of G are singular matrices. An immediate consequence of this result is that if any element of G is a non-singular matrix then all elements of G must be non-singular matrices. (b) Consider the set of singular matrices G of the form x x, (0) x x where x R and x 0. Prove that G is a group with respect to matrix multiplication. Determine the matrix corresponding to the identity element of G. Determine the inverse of the element specified in eq. (0). Observe that x x y y x x y y z z, where z xy. z z This demonstrates that the elements of G satisfy closure on matrix multiplication. Next, we note that ( x x ) x x, x x x x which implies that is the identity element. Finally, ( x x x x e )( ( ) ), () ( ), 5

6 which implies that the group inverse of the element specified in eq. (0) is ) (. () (c) The group defined in part (b) is isomorphic to a well known group. Identify this group. Consider the function from G R that maps the elements x x x, for all x R, (3) x x where R R 0 {0} is the group of non-zero real numbers with respect to multiplication. Thismapisanisomorphism. Itiseasytocheckthatthegroupmultiplicationlawispreserved, since ( x x )( y y ) ( x x y y xy xy ) xy xy (x)(y) xy, is in one-to-one correspondence with multiplication in R. Moreover, the identity [eq. ()] maps to, which is the identity of R. Finally, the inverse given in eq. () is mapped by eq. (3) to /(x), which is the inverse of x in R. We conclude that G R. We can see the isomorphism more explicitly by considering the equivalent representation, x x S S, where S, x x A straightforward computation yields x x x x x Thus, the matrix representation given in eq. (0) is completely reducible and is the direct sum of two one dimensional representations. We can simply discard the zeros, which leaves a one-dimensional representation that is isomorphic to R with the map given by eq. (3). 4. Consider the dihedral group D 4. (a) Write down the group multiplication table. The elements of D 4 are defined by: D 4 {,r,r,r 3,d,rd,r d,r 3 d}, where the elements satisfy the relations, r 4 d and dr r 3 d. (4) We have used the notation e to define the identity element of D 4. 6

7 Using eq. (4), the group multiplication table is immediately obtained: r r r 3 d rd r d r 3 d r r r 3 d rd r d r 3 d r r r r 3 rd r d r 3 d d r r r 3 r r d r 3 d d rd r 3 r 3 r r r 3 d d rd r d d d r 3 d r d rd r 3 r r rd rd d r 3 d r d r r 3 r r d r d rd d r 3 d r r r 3 r 3 d r 3 d r d rd d r 3 r r (b) Enumerate the subgroups, the normal subgroups and the conjugacy classes. There are eight proper subgroups of D 4 : {,r } {,d} {,rd} {,r d} {,r 3 d} Z, {,r,r,r 3 } Z 4, {,r,d,r d} {,r,rd,r 3 d} D. Among these subgroups, four are normal subgroups: 3 {,r } Z, {,r,r,r 3 } Z 4, and {,r,d,r d} {,r,rd,r 3 d} D. Finally, we enumerate the classes: C {}, C {r,r 3 }, C 3 {r }, C 4 {d,r d} and C 5 {rd,r 3 d}. (5) (c) Identify the factor groups. Is the full group the direct product of some of its subgroups? Using the results of part (b), the possible factor groups are: D 4 /Z D, D 4 /Z 4 Z, D 4 /D Z. (6) The last two factor groups are identified uniquely as Z, since this is the only group of two elements. The identification of the first factor group is non-trivial, since there are two possible groups of order four D and Z 4. Note that D is not a cyclic group, whereas Z 4 is a cyclic group. However, it is clear that D 4 /Z is not a cyclic group. In particular, writing out the left cosets, { } D 4 /Z {,r }, {r,r 3 }, {d,r d}, {rd,r 3 d}, 3 One can prove that if a finite group G possesses a subgroup H that contains exactly half the number of elements of G, then H is a normal subgroup of G. 7

8 and identifying {,r } as the identity element of D 4 /Z, it is straightforward to check that the squares of all the other elements of D 4 /Z yields the identity element, which is not in general satisfied by the elements of Z 4. In light of eq. (6), the only possible candidates for writing D 4 as a direct product of its subgroups are Z D or Z Z 4. But the latter two are direct products of abelian groups, which imply that the corresponding direct product groups are abelian, whereas D 4 is a non-abelian group. Hence, D 4 is not a direct product of some of its subgroups. On the other hand, D 4 can be expressed as a semi-direct product of its subgroups in two different ways, D 4 Z4 Z D Z. (7) If we take D {,r,rd,r 3 d}, then we identify Z {,d} in both semi-direct products of eq. (7). 4 Note that D 4 cannot be written as Z D, since the first group of the semi-direct product is the normal subgroup. But, with Z {,r }, we see that one does not obtain all elements of D 4 in the form of g g, with g Z {,r } and g D. 5. The center of a group G, denoted by Z(G), is defined as the set of elements z G that commute with all elements of the group. That is, Z(G) {z G zg gz, g G}. (a) Show that Z(G) is an abelian subgroup of G. To prove that Z(G) is a subgroup of G, we must prove that: (i) z, z Z(G) z z Z(G), (ii) e Z(G), where e is the identity, (iii) z Z(G) z Z(G). To prove (i), we note that z, z Z(G) means that z g gz, for all g G, (8) z g gz, for all g G. (9) Multiply eq. (8) on the right by z to obtain z gz gz z. (0) Then, use eq. (9) to write z gz z z g. Then, eq. (0) can be rewritten as z z g gz z, which means that z z commutes with any element g G. Hence, z z Z(G). 4 If D {,r,d,r d} then we identify Z {,rd} in the second semi-direct product in eq. (7). 8

9 The proof of (ii) is trivial since e commutes with all elements of G. Finally to prove (iii) we note that z Z(G) means that zg gz for all g G. Multiplying this equation on the left by g and on the right by g yields Taking the inverse of eq. () yields g z zg, for all g G. () z g gz, for all g G. Hence, z Z(G). Thus, we have succeeded in showing Z(G) is a subgroup of G. Finally, it should be clear that Z(G) is an abelian subgroup. As previously noted, for any z, z Z(G), eq. (8) is satisfied. In particular, choosing g z in eq. (8), it follows that z z z z. This arguments continues to hold for any choice of z,z Z(G). Thus, we conclude that Z(G) is an abelian subgroup of G. (b) Show that Z(G) is a normal subgroup of G. To show that Z(G) is a normal subgroup, one must show that for any z Z(G) and g G, we have gzg Z(G). By definition, if z Z(G) then gz zg for all g G. Hence, for any z Z(G), we have gzg zgg z Z(G) for all g G, as required for a normal subgroup. (c) Find the center of D 4 and construct the group D 4 /Z(D 4 ). Determine whether the isomorphism D 4 [D4 /Z(D 4 )] Z(D 4 ) is valid. The multiplication table for D 4 was given in part (a) of problem 4. Inspection of the multiplication table reveals that: Z(D 4 ) {e,r } Z, where the identification of the center follows from the fact that any finite group of two elements must be isomorphic to Z. The left cosets of D 4 with respect to the Z subgroup are: Z {e,r }, rz {r,r 3 }, dz {d,r d}, rdz {rd,r 3 d}, which exhausts all the elements of D 4. We identify the quotient group { } D 4 /Z {e,r }, {r,r 3 }, {d,r d}, {rd,r 3 d}. 9

10 From the multiplication table for D 4, one can construct the multiplication table for D 4 /Z, {e,r } {r,r 3 } {d,r d} {rd,r 3 d} {e,r } {e,r } {r,r 3 } {d,r d} {rd,r 3 d} {r,r 3 } {r,r 3 } {e,r } {rd,r 3 d} {d,r d} {d,r d} {d,r d} {rd,r 3 d} {e,r } {r,r 3 } {rd,r 3 d} {rd,r 3 d} {d,r d} {r,r 3 } {e,r } This is clearly not a cyclic group with one generator. Hence, it is not isomorphic to the cyclic group Z 4, which leave only one remaining possibility, D. Indeed, one can check that the multiplication table above is equivalent to that of D. Hence, D 4 /Z D. Finally, if the isomorphism D 4 [D4 /Z(D 4 )] Z(D 4 ) were valid, then D 4? D Z. But this identification is incorrect. In particular, D 4 is a nonabelian group, whereas both D and Z are abelian groups. Thus, it follows that D Z is abelian, which means that this group cannot be isomorphic to the nonabelian group D An automorphism is defined as an isomorphism of a group G onto itself. (a) Show that for any g G, the mapping T g (x) gxg is an automorphism (called an inner automorphism), where x G. To show that T g (x) gxg is an automorphism, we must show that it is a homomorphism from the group G to itself that is one-to-one and onto. To prove that T g is a homomorphism, one must verify that T g (x)t g (y) T g (xy), for all x,y G. () That is, T g (x) preserves the group multiplication table. The computation is straightforward: T g (x)t g (y) (gxg )(gyg ) gxyg T g (xy). To see that T g (x) gxg is one-to-one and onto (i.e. it is an isomorphism), we can invoke the rearrangement lemma. Multiplication on the left and/or on the right by a fixed element of G simply reorders the group multiplication table. 5 Hence, we conclude that T g is an isomorphism from G G. That is, T g is an automorphism of the group G. 5 One can also prove the one-to-one and onto properties directly. To prove that the homomorphism is one-to-one, one must show that T g (x) T g (y) x y. But, T g (x) T g (y) implies that gxg gyg. Multiplying this equation on the left by g and on the right by g then yields x y. To prove that the homomorphism is onto, one must show that for all y G, there exists an x G such that T g (x) y. In this case, it is sufficient to choose x g yg. Evaluating T g (x) for this choice, T g (g yg) g(g yg)g y, as required. Thus, for any choice of y G, we have explicitly determined the required x, namely x g yg, such that T g (x) y. That is, the homomorphism maps G onto itself. 0

11 (b) Show that the set of all inner automorphisms of G, denoted by I(G), is a group. Define I(G) {T g g G}. Since T g is an automorphism, we can introduce a group multiplication law that consists of the composition of two maps. In particular, T g T g (x) T g (g xg ) g g xg g (g g )x(g g ) T g g (x), which holds for any x G. Hence, the composition of two maps is given by: T g T g T g g. (3) It follows that I(G) satisfies the axioms of a group by virtue of the fact that the group G satisfies the group axioms. In particular, eq. (3) implies that I(G) is closed with respect to the group multiplication law. Moreover, associativity is guaranteed because g (g g 3 ) (g g )g 3 implies that T g (T g T g3 ) (T g T g )T g3 T g g g 3. The identity of I(G) is T e (where e is the identity element of the group G) since The inverse of T g is T g, since T g T e T e T g T ge T eg T g. T g T g T g T g T gg T g g T e. Thus, the group axioms are satisfied, which implies that I(G) is a group. (c) Show that I(G) G/Z(G), where Z(G) is the center of G. The kernel of the map f : G G is defined by K ker f {g G f(g) e }, where G is the image of f and e is the identity element of G. Introduce the two homomorphisms, φ : G G/K ψ : G/K G given by φ(g) gk, given by ψ(gk) f(g). It follows that ψ φ(g) f(g). It is straightforward to show that ψ is an isomorphism, in which case we can identify G G/K. (4) This result can be represented diagrammatically by: f G G φ G/K ψ

12 Consider the homomorphism, f : G I(G), given by f(g) T g. Note that f is onto, i.e. I(G) is the image of f. The kernel of f is K {g G} f(g) T e }, where T e is the identity element of I(G), i.e. T e (x) x. Thus, K consists of all elements of G satisfying T g T e, or equivalently, gxg x, which implies that gx xg for all x G. We recognize this as the center of G, denoted by Z(G) in problem 4. Using eq. (4), it follows that I(G) G/Z(G). (5) (d) Show that the set of all automorphisms of G, denoted by A(G), is a group and that I(G) is a normal subgroup. (The factor group A(G)/I(G) is called the group of outer automorphisms of G.) Let A(G) be the set of all automorphisms of G. To show that this is a group, we must define the group multiplication law. As in the case of part (b), we define A A (g) A (A (g)), for A,A A and g G. That is the multiplication law is simply the composition of maps. It is straightforward to verify that the group axioms are satisfied. Note that since an automorphism is one-to-one and onto, each element of A(G) possesses a unique inverse. Next, we demonstrate that the set of inner automorphisms, {T g g G}, is a normal subgroup of A(G). To do this, one must show that AT g A I(G), for all A A(G). Consider, AT g A (x) AT g (A (x)) A(gA (x)g ) A(g)A(A (x))a(g ) A(g)xA (g) T A(g) (x), (6) where we have used the fact that A is a homomorphism, which therefore satisfies A(g g ) A(g )A(g ) and A(g ) A (g), for any g,g,g G. (7) It follows that AT g A T A(g) I(G). 7. Consider an arbitrary orthogonal matrix R, which satisfies RR T (where is the identity matrix). (a) Prove that the possible values of detr are ±. Using the fact that detr T detr, it follows that det(rr T ) (detr)(detr T ) [detr], (8) since RR T implies that det(rr T ) det. Taking the square root of eq. (8) yields detr ±.

13 (b) The group SO() consists of all orthogonal matrices with unit determinant. Prove that SO() is an abelian group. Suppose that Q SO(). If we parameterize a b Q, c d then we can find relations among the parameters a, b, c and d by imposing the conditions Q T Q and detq. That is, a c a b a +c ab+cd 0 b d c d ab+cd b +d, 0 and det Q ad bc. Hence, the relations among the parameters a, b, c and d are determined by the following conditions, a +c b +d, ab+cd 0, ad bc. (9) We now consider two cases. First if c 0, it follows that d ab/c. Inserting this result back into eq. (9) yields ad bc a b c bc b ( a +c ) b c c, after using eq. (9). That is, c b. It immediately follows that d ab/c a, and we conclude that the most general SO() matrix is given by a b Q. b a In light of eq. (9), c b yields a +b, which implies that a,b. Thus, it is convenient to parameterize a and b by defining a cosθ and b sinθ. Hence, the most general SO() matrix is given by cosθ sinθ Q, (30) sinθ cosθ where 0 θ < π. Next, we examine the case of c 0. In this case, eq. (9) yields a, ab 0, and ad. It follows that b 0 and a d ±. Hence the form for Q in this case (where a d ± and b c 0) is consistent with eq. (30). It is now a simple matter to show that SO() is a group and any two elements of SO() of the form given in eq. (30) commute. In particular, cosθ sinθ cosθ sinθ sinθ cosθ sinθ cosθ cosθ cosθ sinθ sinθ sinθ cosθ +cosθ sinθ sinθ cosθ cosθ sinθ cosθ cosθ sinθ sinθ cos(θ +θ ) sin(θ +θ. (3) sin(θ +θ ) cos(θ +θ ) 3

14 The form of the group multiplication law given above exhibits closure. The identity corresponds to taking θ 0 in eq. (30), and the inverse of Q is obtained by taking θ θ. Finally, if we interchange θ and θ in eq. (3), we recover the same result. Hence, all products of SO() elements are commutative (which implies associativity), and we conclude that SO() is an abelian group. (c) The group O() consists of all orthogonal matrices, with no restriction on the sign of its determinant. Is O() abelian or non-abelian? (If the latter, exhibit two O() matrices that do not commute.) The matrix Q given in eq. (30) is also an element of O(). An element of O() that is not an element of SO() is 0. 0 But this matrix does not commute with Q. In particular, cosθ sinθ 0 cosθ sinθ, sinθ cosθ 0 sinθ cosθ whereas 0 cosθ sinθ cosθ sinθ. 0 sinθ cosθ sinθ cosθ Hence, we conclude that O() is a non-abelian group. REMARK: Note that SO() is a normal subgroup of O(). To prove this result, consider the homomorphism, f : O() {+, }, which is defined by f(a) deta, for A O(). The kernel of f is SO(), since the latter corresponds to the set of all elements of O() with determinant equal to one. Hence, O()/kef f {+, }. Since we can identify Z {+, } where the group operation is ordinary multiplication, we can conclude that O()/SO() Z. However, it does not follow that O() SO() Z. Indeed, O() is a nonabelian group whereas SO() Z is an abelian group. Nevertheless, it is true that O() is a semi-direct product, O() SO() Z. To show this, we simply need to exhibit a Z subgroup of O() such that SO() Z {e}, where e is the identity element of O(). A possible choice for the Z subgroup of O() that satisfies this requirement is, Z { 0, 0 } 0. 0 One can easily verify that this Z subgroup is not a normal subgroup of O(). In particular, gz g Z for all g O(), as one can easily check. 4